Methods of predicting vehicle movements on roads by computer simulations mainly include macroscopic methods and microscopic methods.
In the macroscopic methods, the time required for vehicles to pass through roads, a presence or an absence of congestion, and so forth are predicted from a relationship between a traffic volume (number of vehicles that can pass through per unit time) determined by speed limits, numbers of lanes, road widths, and so forth and traffic demand (number of vehicle for which usage is expected per unit time). The macroscopic methods that predict vehicle movements with such an approach are capable of predicting traffic conditions in a wide area in a short calculation time.
However, because the macroscopic methods do not separately track the behaviors of individual vehicles, it is not possible to focus on a specific vehicle and to predict traveling time required for the traveling distance, etc. for each vehicle. In addition, because simulations are performed assuming that vehicle densities on roads are uniform, there is a problem in that an increase and a decrease in vehicle density occurring in the rear affects traveling velocities of vehicles in the front.
In contrast, because the microscopic methods separately track the behaviors of individual vehicles, the above-described problem that traveling time cannot be separately predicted for the individual vehicles is eliminated. In addition, in the microscopic methods, because vehicle velocities for the individual vehicles are calculated from situations in the surroundings, such as the distance to an intersection, the behaviors of vehicles in front, and so forth, the problem that an increase and a decrease in the vehicle density occurring in the rear affects traveling velocities of vehicles in the front is eliminated.
However, because the microscopic methods separately analyze the behaviors of individual vehicles, the amount of calculation as a whole is large, and there is a problem in that a long calculation time is required.
Therefore, Patent Literature 1 discloses a traffic-flow simulator in which links, which are road models simulating roads, are divided into a plurality of blocks, and vehicle movements are calculated for each block, thereby improving the simulation speed.